Dendrobium huoshanense plant stems were collected from the Huoshan County, Anhui Province of China. Referring to the subject group’s pre-extraction method (Ye et al., 2022), the powder was degreased with 95% ethanol at 25°C for 24 h, centrifuged, and retained. Then, the Sevage method was used to deproteinize the precipitates after extraction with distilled water. Impurities with molecular weights less than 3,500 Da were removed by dialysis. Eventually, DHP was obtained by freeze-drying. The structure of DHP was resolved using SEM, FT-IR, HPGPC, and ion chromatography (IC). Specific experimental steps are described in Supplementary Material S1.

C57BL/6J mice aged five to 8 weeks were maintained in cages under standard laboratory conditions (20°C–25°C temperature, 40%–70% relative humidity). The mice were fed a high-fat diet (fat, 60%; carbohydrates, 20%; protein, 20%; Jiangsu Xietong Pharmaceutical Bio-engineering Co., Ltd) and high-fructose drinking solution (10% D-fructose, Shanghai Macklin Biochemical Co., Ltd) diet for 12 weeks to establish the NAFLD mouse model.

As illustrated in Figure 1, the mice were allocated into six distinct groups (n = 6 mice per group) in a random manner: the control group (CON), the model group (MOD), the pioglitazone group (PGZ, 10 mg/kg), the DHP-low dose group (DHP-L, 50 mg/kg), the DHP-medium dose group (DHP-M, 100 mg/kg), and the DHP-high dose group (DHP-H, 200 mg/kg). All mice, except the CON, were fed High-Fat and High-Fructose drinking (HFHF) and administered the various drugs in the period from weeks 5–12.

At the 4 weeks feeding time point, mice in DHP-L, DHP-M and DHP-H groups were injected with 50 mg/kg, 100 mg/kg and 200 mg/kg of DHP by gavage once a day, respectively, mice in the PGZ group were given with 10 mg/kg PGZ per day and MOD mice were given the same dose of the saline for 8 weeks. CON mice were fed standard chow (fat, 11.1%; carbohydrates, 67.4%; protein, 21.5%; Jiangsu Xietong Pharmaceutical Bio-engineering Co., Ltd 3) and normal water and were given with saline after the fourth week according to their body weight. Before euthanasia, the mice were fasted overnight. Euthanasia was performed by intraperitoneal injection of 3% pentobarbital. Blood was then collected from the abdominal aorta, centrifuged at 1,500 rpm, 4°C, for 15 min, and the supernatant was collected. The mouse livers were collected in lyophilized tubes and stored at −80°C. This study was approved by the Ethics Committee of the Anhui University of Chinese Medicine.

The body weight of the mice were recorded on a weekly basis. The liver index was calculated using the following formula: Liver index = liver weight (mg)/body weight g).

Liver tissues were prepared in 4 μm tissue sections by conventional paraffin embedding, stained according with hematoxylin and eosin (H&E) and Oil Red O, viewed under a microscope, and photographed.

The levels of TC, TG, ALT, AST, low-density lipoprotein cholesterol (LDL-C), and high-density lipoprotein cholesterol (HDL-C) in the serum were calculated using a fully automated biochemical analyzer (Chemray-240).

Liver tissues were homogenized in RIPA buffer. Then, the sample was vortexed and centrifuged at 2000 rpm for 5 min at 4°C. The supernatant was collected and stored at 80°C before performing ELISA. TC, TG, TNF-α, IL-1β, and IL-6 were detected in mouse livers using the corresponding kits (ColorfulGene Biological Technology Co.,Ltd (Wuhan, China) according to the manufacturer’s instructions.

Liver tissue (30 mg) was placed in 400 μL of methanol-water (V: V = 7:3) solution, ground, and centrifuged. Then, the supernatant was collected for LC-MS analysis. The separation was carried out on the Waters Acquity UPLC HSS T3 C18 column (2.1 *100 mm, 1.8 µm) at 40°C with mobile phases A and B consisting of water and acetonitrile with 0.1% formic acid, respectively. The gradient elution procedure was as follows: within 12 min, mobile phase A was reduced from 95% to 10%, held for 1 min at 10%, and then returned to 95% for 3 min. For mass spectrometry measurements, an Agilent Q-TOF-6545 mass spectrometer was used with an interface voltage of 2.5 kV for positive ions and −1.5 kV for negative ions.

For analysis of metabolites, m/z values, formulas, and MS/MS fragmentations were compared with online databases, including Kyoto Encyclopedia of Genes and Genomes (KEGG) database (http://www.kegg.jp/), the Human Metabolome Database (HMDB) (http://www.hmdb.ca), and PubChem (https://puccbi.nlm.nih.gov/), among others. MetaboAnalyst 5.0 (https://www.metaboanalyst.ca/MetaboAnalyst/) was applied to conduct an orthogonal partial least squares discriminant analysis (OPLS-DA), partial least squares discriminant analysis (PLS-DA), and principal component analysis (PCA). Using MetaboAnalyst 5.0, we screened differential metabolites by the variable importance in projection (VIP) > 1 and the p < 0.05, followed by metabolic pathway analysis.

Total RNA was extracted from liver tissue and reverse-transcribed into cDNA with the SPARK script II RT Plus Kit (SparkJade Co., Ltd. Shandong, China), using Sangon Biotech synthesized primers (Sangon Biotech Co., Ltd. Shanghai, China) (Table 1). SYBR Green in a Light Cycler@480 System (Roche, Switzerland) was used to perform qRT-PCR analysis.

Liver tissues from the mice ×were fixed in formalin solution, cut into paraffin sections, deparaffinized, and hydrated for antigenic repair. F4/80, CD68, and p-NF-κB primary antibodies (1:200) were incubated overnight at 4°C before being washed with PBS (5 min × 3 times) (Cell Signaling Technology, USA). Following incubation with the primary antibodies, a secondary antibody (1:300) was added, followed by incubation for 1 h at room temperature. Sections were then washed with PBS (5 min × 3 times). In the final step of the process, after the surface of the tablet had been treated with a quenching agent that inhibits fluorescence, the tablet was sealed. Staining was visualized and photographed using a fluorescence microscope (Leica, Germany).

Fresh liver tissues were acquired and ground, then, RIPA buffer was added for further grinding. Next, bicinchoninic acid assay was performed to determine the protein concentration in the supernatant after centrifugation of the total protein extract. SDS-PAGE gels were prepared. After sampling, electrophoresis and membrane transfer, then, was performed; primary antibodies (TLR4 (ZENBIO, China), MyD88 (Proteintech, China), NF-Κb (Proteintech, China), p-NF-κB, and β-actin (ZENBIO, China) were used at the appropriate dilution, followed by incubation overnight at 4°C in the refrigerator. A secondary antibody solution was added to the strips, and they were incubated at room temperature for 2 h. Enhanced chemiluminescence (ECL) reagents were used to detect antigen-antibody complexes, visualized using an ECL chemiluminescence imaging system (G&E Healthcare, USA). Grayscale analysis was conducted using Image Lab.

All experimental results were expressed as the means ± standard deviation (mean ± SD). GraphPad Prism 9.0 was used to analyze the data. The difference between multiple groups was compared using a one-way ANOVA (One-way ANOVA), with p < 0.05 indicating statistical significance.

The molecular morphology of DHP was examined using a scanning electron microscope and observed at image magnifications of 1.00 k×, 500.00 ×, and 200.00 ×. Microstructural analysis of DHP revealed an irregular lamellar structure with an uneven surface (Figure 2A). As shown in Figure 2B, the DHP exhibited strong absorption peaks at 3,398 cm^-1 and 2,887 cm^-1, which were caused by stretching vibration of the O–H and the C–H bond, respectively, and were polysaccharides-specific peaks. The characteristic absorption peaks at 951 cm^-1 and 896 cm^-1 indicates that the DHP has both the typical α-configuration and β-configuration. As shown in Figure 2A-C, the weight and polydispersity index (Mw/Mn) and number-average molecular weight (Mw and Mn) of DHP were determined using SEC-MALLS-RI, as 179.935 kDa, 1.224 and 146.96 kDa, 1.224, respectively. The purity of DHP was 89.55%. Being a heteropolysaccharide composed of Man and Glc, DHP was found to consists of 68.97% and 31.03% mannose and glycoside, respectively.

As shown in Figure 3A, the MOD group achieved a significantly higher body weight than the CON group over the entire experimental period (p < 0.01). Compared with that in the MOD group, DHP administration resulted in a significant decrease in body weight (p < 0.01). To determine whether the DHP-induced weight decrease was due to reduced food intake, the food intake of mice in the MOD group and the DHP-L, DHP-M, and DHP-H groups was measured. As illustrated in Supplementary Figure S2A, the average daily food intake of mice in the MOD and DHP-L, DHP-M, and DHP-H groups was similar. The increased liver weight and liver index in MOD group were remarkably decreased (p < 0.05) following administration of DHP and PGZ (Figures 3B, C).

According to Figures 3D, E, the lobule structure of the liver lobules of mice in the CON group was clearly discernible, and the structure of the liver tissue was normal. The MOD group showed obvious fatty degeneration of hepatocyte tissue and excessive accumulation of lipid droplets and accompanied by the infiltration of inflammatory cells, suggesting that HFHF might cause severe liver injury in mice. Compared with MOD, DHP significantly reduced the fatty degeneration of hepatocytes, the infiltration of inflammatory cells, and local necrosis, which suggests that DHP ameliorated HFHF-induced NAFLD in the present mouse model.

Serum TG, TC, ALT, AST, LDL-C levels significantly increased while those of HDL-C levels obviously decreased in the MOD group compared with those in CON group (Figures 4A–F, p < 0.01). A similar trend was noticed for TG and TC levels in the liver, with the MOD group showing a higher level of these metabolites (Figures 4A–F, p < 0.01). Compared with those in the MOD group, the serum TG, TC, ALT, AST, and LDL-C levels were significantly decreased and those of HDL-C levels obviously increased following DHP or PGZ treatment; furthermore, the liver TG and TC levels were considerably reduced after DHP or PGZ treatment (Figures 4A–F, p < 0.01). These results indicate that DHP protects NAFLD mice from hepatic injury.

As illustrated in Figures 5A–C, liver TNF-α, IL-6, and IL-1β levels in the mice fed HFHF remarkably increased (p < 0.01) compared with those in the CON group. Treatment with DHP-H effectively inhibited liver IL-1β, IL-6, and TNF-α expression compared with that in the MOD group. Compared with the CON group, the mRNA levels of TNF-α, IL-6, and IL-1β in the liver were elevated in the MOD group (p < 0.01). In comparison with the MOD group, DHP administration significantly reduced hepatic mRNA levels of IL-6, IL-1β, and TNF-α (p < 0.01) (Figure 5D–F).

Immunohistochemical staining of F4/80 and CD68 showed that DHP reduced macrophage infiltration in the liver, further supporting the results of ELISA and PCR (Figure 6A, B). Furthermore, DHP treatment resulted in a significantly decreased in the expression of TLR4, Myd88, and p-NF-κB (Figure 6C). The results of the immunostructures were consistent with those of the Western blot (Figure 6D–E). These results show that DHP ameliorated HFHF diet-induced hepatic inflammation by modulating the TLR4/NF-κB signaling pathways.

A further metabolomics study was conducted in order to determine the effect of DHP on liver metabolic responses in NAFLD mice. The liver metabolic profiles were obtained using UHPLC-MS/MS. In the metabolomics study, 4,146 peaks were identified (the counts of ESI + ions were 2,859 and 1,287, respectively). The PCA and PLS-DA results indicated distinguishable liver metabolites among the CON, MOD, and DHP groups (Figures 7A, B). According to the PCA and PLS-DA score plots, good aggregation was observed among the CON, MOD, and DHP groups, indicating good reliability of the method. Furthermore, as shown in Figure 7C, F, supervised OPLS-DA was used to compare the MOD and DHP groups in order to better distinguish between these groups and improve the effectiveness and analytical capability of the model. By examining the OPLS-DA score plot, it was demonstrated that there was a significant difference between the MOD and DHP groups in the positive/negative model, indicating that the metabolites were different. These results indicate that DHP can regulate metabolic disorders caused by feeding HFHF in the present mouse model.

Based on the positive/negative ion mode, the R2 and Q2 results were (0.997, −0.013) (Figure 7D) and (0.991, −0.079) (Figure 7G). The results showed that the model was reliable and had predictive ability. Based on the volcano plot, we found that compared with the model group, 165 metabolites were dramatically decreased and 278 metabolites were significantly increased in the positive ion mode after DHP treatment; and 53 metabolites were greatly decreased and 107 metabolites were significantly increased in negative ion mode following DHP treatment (Figures 7E–H).

Using untargeted metabolomics, 49 differential liver in the liver were determined between the MOD groups and the DHP groups. These metabolites were identified by screening with the following screening criteria: p < 0.05, VIP >1, and KEGG database searches (Table 2). Compared with the MOD group, 19 metabolites were obviously decreased and 30 metabolites were significantly increased in the DHP group (Figure 8A). The main categories for these 49 differential metabolites were as follows: amino acids and its metabolites (18.36%); organic acids and its derivatives (14.28%); aldehydes, ketones, esters (10.20%); and fatty acids (8.16%) (Figure 8B). MetabolAnalyst 5.0 was used to performed KEGG-based metabolic pathway enrichment analyses of various liver metabolites to identify the major metabolic pathways affected in the MOD and DHP groups. Results showed that these different metabolites primarily enriched in glycine, serine, and threonine, nicotinate and nicotinamide metabolism, and arachidonic acid metabolism, among others (Figures 8C, D).